Best AI research tools by task
Honest category maps of the tools researchers actually use, including the ones that compete with ours. Every roundup lists tools alphabetically rather than ranked, states what each tool cannot do, and links to the dated source each claim came from — our full comparison, or the maker’s own page.
- Best AI research gap finders
Nine AI tools researchers use to find open research questions — what each is best at, where each falls short, and how to pick. Updated September 2026.
9 tools · AnswerThis, Connected Papers, Consensus, Elicit, ResearchRabbit, Science AI Journal, SciSpace, scite, Semantic Scholar
- Best AI peer review tools
Six ways to get a manuscript reviewed before submission — AI reviewers, open review, and our own. What each is best at and where each falls short.
6 tools · OpenReview, Paperpal, PeerGenius, Reviewer3, ScholarsReview, Science AI Journal
- Best journal finders
Fourteen tools for deciding where to submit your paper — free and paid, one publisher or all. What each takes in, what it returns, where it falls short.
14 tools · B!SON, Clarivate Manuscript Matcher, Editage Journal Finder, Elsevier Journal Finder, JANE, Journal Suggester (journalsuggester.com), JournalGuide, PubScope, Science AI Journal, Scientific Journal Finder, Springer Nature Journal Finder, Taylor & Francis Journal Suggester, Trinka Journal Finder, Wiley Journal Finder
- Best free citation generators
Five ways to turn a reference into a formatted citation — managers, student generators, and a DOI-first converter. What each is best at, and its limits.
5 tools · BibGuru, Mendeley, Science AI Journal, Scribbr, Zotero
Why we publish these
Science AI Journal appears in these roundups alongside its competitors. We wrote them, so read them sceptically — but the structure is deliberately hard to rig: alphabetical ordering, a required limitation on every entry including ours, no statistics about anyone else’s product, and a link from each entry to the dated source its claims came from — a comparison page we wrote, or the maker’s own page. Our own measured performance, including the runs that failed, is on our benchmarks page.